Elderly people (age 65 years and older) are the most rapidly growing segment of the US population and the incidence of epilepsy is greatest in this age group. Little information is available about the use or outcomes associated with antiepileptic drugs (AEDs) by these vulnerable persons. We will examine pattems of AED use and associated outcomes among nursing home (NH) and community based elderly. For the proposed research, we have available two large electronic data sets: (1) the Beverly Enterprises Inc. centralized records (the Minimum Data Set and physicians' orders) for two calendar years (1999, 2000) for all residents (N~140,000 for two years) living in 501 NHs in 31 states, and (2) the Health Care Financing Administration's (HCFA) integration of the Medicare Current Beneficiaries Survey with Medicare claims files (N~12,000 elderly per year) for 7 calendar years. Both of these data sets have been used in part during the current project (1997-2002) by the same multidisciplinary team that will be working in the proposed project. We have a thorough working knowledge of the complexities of these data sets and the richness of information they contain. In our current research, we found a prevalence of 7.7% for AED use by NH admissions. Significant covadates with AED use at admission were epilepsy/seizure, bipolar depression, age group, and cognitive performance. During the first 3 months following admission, an additional 2.7% people were initiated on AEDs. Epilepsy/seizure indication was present in 5.8% of all admissions and an additional 1.47% of the follow-up cohort during the first 3 months. Specific aims of the proposed project are: (1) outcomes associated with any AED use by eldedy, (2) pharmacoepidemiology, characteristics, and outcomes associated with CBZ, PHT and VPA, and (3) the pharmacoepidemiology of other specific AEDs. Patient outcomes will be captured through incidence events, while AED use can be tracked continuously over time. Multivariate-modeling techniques will be used, controlling for demographics, co-morbidities, and regional effects. Other P50 teams are studying the elimination kinetics of AEDs, and we will use this knowledge to develop age, gender and co-medication specific estimated total and unbound concentrations for NH residents.